Using sketches and knowledge bases for geo-spatial image retrieval

This paper presents research in the field of knowledge management for geo-spatial imagery including scanned aerial photos and satellite images. We have developed a web-based system that allows users to query a database of images not only using metadata, but also drawing sketches of configurations of objects they are interested in as well as inputting textual descriptions of their intended task. Our system integrates case-based reasoning techniques to form a knowledge base from previously issued queries that can be exploited to improve future query processing and to build organizational memory through experience capture. The effective design and implementation of a user-friendly graphic user interface plays an important role for the system to provide improved human–computer interaction and decision support.